Apply function 'str' object has no attribute 'str'-CodePudding

cause I don't really understand why it works in one format and doesn't work in another.

Works:

df['team'] = df['team'].str.extract(r'(\w ) ')

Doesn't work:

def clear_teams(gr):
    return gr.str.extract(r'(\w ) ')
        
df['team'] = df['team'].apply(clear_teams)

I recive an error:

AttributeError: 'str' object has no attribute 'str'

Why it doesn't work, can someone explain it to me ? pleas :) How it has str attribute one time and another doesn't ....

CodePudding user response：

If use Series.apply then in function gr is scalar, function loop by element of Series. So cannot use Series functions for it like str.extract.

def clear_teams(gr):
    print (type(gr))
    print (gr)

    return gr.str.extract(r'(\w ) ')

If use Series.pipe then gr is Series, co all working correct:

df['team'] = df['team'].pipe(clear_teams)

Or:

df['team'] = clear_teams(df['team'])

CodePudding user response：

When you use apply function on a pd.Series, the referenced function gets the actual cell value for each value:

def clear_teams(gr):
    # if df['team'] dtype is str, than gr is string, 'str' attribute is a 'pd.Series` attribute
    return gr.extract(r'(\w ) ')
        
df['team'] = df['team'].apply(clear_teams)